@kensio/pinyinjs 1.7.10 → 1.7.12

package/dist/decode/de-rule.d.ts

@@ -60,10 +60,10 @@ export declare const POTENTIAL_DE: EdgeRule;
  * Only a key **no source has tagged** is forbidden, which is the same line
  * {@link TEACHING_JIAO} draws for a word ending in 教. 的确, 的士 and 的哥 are
  * words jieba counted and tagged, and every one of them can genuinely begin
- * where this fires: 我的确知道, 我要一辆的士, 那个的哥. That leaves 我的哥哥们
- * reading `wǒ dī gēgē men`, since 的哥 is a taxi driver and jieba has counted
- * them. What this does reach is the untagged tail — a reading somebody asserted
- * rather than a word anybody counted — where the 的 is the particle every time.
+ * where this fires: 我的确知道, 我要一辆的士, 那个的哥. What it does reach is
+ * the untagged tail — a reading somebody asserted rather than a word anybody
+ * counted — where the 的 is the particle every time. The tagged taxi keys are
+ * {@link import("./taxi-rule.js").TAXI_DI}'s, which is what settles 我的哥哥们.
  *
  * Measured over the 88,866 lines of Tatoeba and zh.wikipedia it forbids edges
  * in 40 runs: 的筆 13, 的這 11, 的真 6, 的對 4, 的卡 4 and 的歷 once. Every one

package/dist/decode/de-rule.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"de-rule.d.ts","sourceRoot":"","sources":["../../src/decode/de-rule.ts"],"names":[],"mappings":"AAUA,OAAO,EAEL,KAAK,QAAQ,EAKd,MAAM,YAAY,CAAC;AAgCpB;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,QAAQ,EAAE,QActB,CAAC;AAkDF;;;;;;;;;;;;;;;;;;GAkBG;AACH,eAAO,MAAM,YAAY,EAAE,QAoB1B,CAAC;AAWF;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,eAAO,MAAM,WAAW,EAAE,QAczB,CAAC"}
+{"version":3,"file":"de-rule.d.ts","sourceRoot":"","sources":["../../src/decode/de-rule.ts"],"names":[],"mappings":"AASA,OAAO,EAEL,KAAK,QAAQ,EAKd,MAAM,YAAY,CAAC;AAgCpB;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,QAAQ,EAAE,QActB,CAAC;AA8CF;;;;;;;;;;;;;;;;;;GAkBG;AACH,eAAO,MAAM,YAAY,EAAE,QAoB1B,CAAC;AAWF;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,eAAO,MAAM,WAAW,EAAE,QAczB,CAAC"}

package/dist/decode/de-rule.js

@@ -1,13 +1,4 @@
-/**
- * The two rules about 得 and 的, which are one character each and three jobs.
- *
- * 得 is stored as the structural particle with `dé` and `děi` as alternates,
- * and 的 is stored as the particle full stop. Both are among the commonest
- * characters in the language, so the words they begin and end are numerous and
- * mostly wrong.
- */
-import { toCharacters } from "../script/characters.js";
-import { tagOf, wordEndingAt, wordStartingAt, wordsStartingAt, } from "./rules.js";
+import { startsLongerWord, tagOf, wordEndingAt, wordStartingAt, } from "./rules.js";
 /**
  * Whether a syllable is `děi`, the modal reading of 得.
  *
@@ -112,8 +103,7 @@ function isDe(syllable) {
  * own and a test that counted those would answer yes everywhere.
  */
 function isSwallowed(context, at) {
-    const longer = (from) => wordsStartingAt(context, from).some((word) => toCharacters(word).length > 1);
-    return longer(at) && !longer(at + 1);
+    return startsLongerWord(context, at) && !startsLongerWord(context, at + 1);
 }
 /**
  * A verb and a 得 in front of a complement is not a word reading `dé`.
@@ -173,10 +163,10 @@ const MODIFIED_TAGS = /^(?:r|n|a|v|b|z)/u;
  * Only a key **no source has tagged** is forbidden, which is the same line
  * {@link TEACHING_JIAO} draws for a word ending in 教. 的确, 的士 and 的哥 are
  * words jieba counted and tagged, and every one of them can genuinely begin
- * where this fires: 我的确知道, 我要一辆的士, 那个的哥. That leaves 我的哥哥们
- * reading `wǒ dī gēgē men`, since 的哥 is a taxi driver and jieba has counted
- * them. What this does reach is the untagged tail — a reading somebody asserted
- * rather than a word anybody counted — where the 的 is the particle every time.
+ * where this fires: 我的确知道, 我要一辆的士, 那个的哥. What it does reach is
+ * the untagged tail — a reading somebody asserted rather than a word anybody
+ * counted — where the 的 is the particle every time. The tagged taxi keys are
+ * {@link import("./taxi-rule.js").TAXI_DI}'s, which is what settles 我的哥哥们.
  *
  * Measured over the 88,866 lines of Tatoeba and zh.wikipedia it forbids edges
  * in 40 runs: 的筆 13, 的這 11, 的真 6, 的對 4, 的卡 4 and 的歷 once. Every one

package/dist/decode/de-rule.js.map

@@ -1 +1 @@
-{"version":3,"file":"de-rule.js","sourceRoot":"","sources":["../../src/decode/de-rule.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AACH,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAEvD,OAAO,EAGL,KAAK,EACL,YAAY,EACZ,cAAc,EACd,eAAe,GAChB,MAAM,YAAY,CAAC;AAEpB;;;;;GAKG;AACH,SAAS,KAAK,CAAC,QAA8B;IAC3C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,IAAI,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC5E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9C;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,eAAe,CAAC;AAEtC;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa;IAChC,IAAI,EAAE,UAAU;IAChB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACjD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;QAC/C,OAAO,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC7C,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACzC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;;;;;;;GAYG;AACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAElD;;GAEG;AACH,MAAM,QAAQ,GAAG,GAAG,CAAC;AAErB;;GAEG;AACH,SAAS,IAAI,CAAC,QAA8B;IAC1C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,GAAG,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC3E,CAAC;AACJ,CAAC;AAED;;;;;;;;;;GAUG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,MAAM,MAAM,GAAG,CAAC,IAAY,EAAW,EAAE,CACvC,eAAe,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,IAAI,CACjC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CACxC,CAAC;IACJ,OAAO,MAAM,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;AACvC,CAAC;AAED;;;;;;;;;;;;;;;;;;GAkBG;AACH,MAAM,CAAC,MAAM,YAAY,GAAa;IACpC,IAAI,EAAE,cAAc;IACpB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,KAAK,CAAC;YACzB,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG;YACxB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,CAAC;YACzB,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EACtB,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAChE,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,WAAW,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC;YAC/C,CAAC,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC;YAC9B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,mBAAmB,CAAC;AAE1C;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,MAAM,CAAC,MAAM,WAAW,GAAa;IACnC,IAAI,EAAE,aAAa;IACnB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,YAAY,KAAK,EAAE;YACxB,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAC1B,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,OAAO,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IACxE,CAAC;CACF,CAAC"}
+{"version":3,"file":"de-rule.js","sourceRoot":"","sources":["../../src/decode/de-rule.ts"],"names":[],"mappings":"AASA,OAAO,EAGL,gBAAgB,EAChB,KAAK,EACL,YAAY,EACZ,cAAc,GACf,MAAM,YAAY,CAAC;AAEpB;;;;;GAKG;AACH,SAAS,KAAK,CAAC,QAA8B;IAC3C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,IAAI,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC5E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9C;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,eAAe,CAAC;AAEtC;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa;IAChC,IAAI,EAAE,UAAU;IAChB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACjD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;QAC/C,OAAO,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC7C,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACzC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;;;;;;;GAYG;AACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAElD;;GAEG;AACH,MAAM,QAAQ,GAAG,GAAG,CAAC;AAErB;;GAEG;AACH,SAAS,IAAI,CAAC,QAA8B;IAC1C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,GAAG,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC3E,CAAC;AACJ,CAAC;AAED;;;;;;;;;;GAUG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,OAAO,gBAAgB,CAAC,OAAO,EAAE,EAAE,CAAC,IAAI,CAAC,gBAAgB,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;AAC7E,CAAC;AAED;;;;;;;;;;;;;;;;;;GAkBG;AACH,MAAM,CAAC,MAAM,YAAY,GAAa;IACpC,IAAI,EAAE,cAAc;IACpB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,KAAK,CAAC;YACzB,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG;YACxB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,CAAC;YACzB,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EACtB,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAChE,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,WAAW,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC;YAC/C,CAAC,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC;YAC9B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,mBAAmB,CAAC;AAE1C;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,MAAM,CAAC,MAAM,WAAW,GAAa;IACnC,IAAI,EAAE,aAAa;IACnB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,YAAY,KAAK,EAAE;YACxB,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAC1B,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,OAAO,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IACxE,CAAC;CACF,CAAC"}

package/dist/decode/edge-context.d.ts

@@ -1,9 +1,3 @@
-/**
- * What a rule can ask about the words around an edge.
- *
- * The lookups every edge rule shares: which word ends here, which starts here,
- * and what part of speech the dictionary gives it.
- */
 import type { EdgeContext } from "./rules.js";
 /**
  * The word ending at a position, longest first, or undefined.
@@ -41,6 +35,20 @@ export declare function wordsStartingAt(context: EdgeContext, at: number, longes
  * The part of speech the dictionary gives a word, or the empty string.
  */
 export declare function tagOf(context: EdgeContext, word: string | undefined): string;
+/**
+ * Whether a word of more than one character starts at a position.
+ *
+ * Every character is a word of its own, so a test that counted those would
+ * answer yes everywhere. What a rule asking this wants to know is whether some
+ * longer word has a claim on the character.
+ */
+export declare function startsLongerWord(context: EdgeContext, at: number): boolean;
+/**
+ * Whether a word of more than one character ends at a position.
+ *
+ * The mirror of {@link startsLongerWord}, and asked for the same reason.
+ */
+export declare function endsLongerWord(context: EdgeContext, at: number): boolean;
 /**
  * What follows a 教 that is teaching where no object does: aspect and the
  * complement marker, which only a verb takes.

package/dist/decode/edge-context.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"edge-context.d.ts","sourceRoot":"","sources":["../../src/decode/edge-context.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AACH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAE9C;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,MAAM,GAAG,SAAS,CAYpB;AAED;;;;;;;;;GASG;AACH,wBAAgB,aAAa,CAC3B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,SAAS,MAAM,EAAE,CAanB;AAED;;GAEG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,MAAM,GAAG,SAAS,CAUpB;AAED;;;;;;GAMG;AACH,wBAAgB,eAAe,CAC7B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,SAAS,MAAM,EAAE,CASnB;AAED;;GAEG;AACH,wBAAgB,KAAK,CAAC,OAAO,EAAE,WAAW,EAAE,IAAI,EAAE,MAAM,GAAG,SAAS,GAAG,MAAM,CAI5E;AAED;;;;;;;GAOG;AACH,eAAO,MAAM,MAAM,aAA0C,CAAC;AAE9D;;GAEG;AACH,eAAO,MAAM,YAAY,MAAM,CAAC"}
+{"version":3,"file":"edge-context.d.ts","sourceRoot":"","sources":["../../src/decode/edge-context.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAE9C;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,MAAM,GAAG,SAAS,CAYpB;AAED;;;;;;;;;GASG;AACH,wBAAgB,aAAa,CAC3B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,SAAS,MAAM,EAAE,CAanB;AAED;;GAEG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,MAAM,GAAG,SAAS,CAUpB;AAED;;;;;;GAMG;AACH,wBAAgB,eAAe,CAC7B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,SAAS,MAAM,EAAE,CASnB;AAED;;GAEG;AACH,wBAAgB,KAAK,CAAC,OAAO,EAAE,WAAW,EAAE,IAAI,EAAE,MAAM,GAAG,SAAS,GAAG,MAAM,CAI5E;AAED;;;;;;GAMG;AACH,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,WAAW,EAAE,EAAE,EAAE,MAAM,GAAG,OAAO,CAI1E;AAED;;;;GAIG;AACH,wBAAgB,cAAc,CAAC,OAAO,EAAE,WAAW,EAAE,EAAE,EAAE,MAAM,GAAG,OAAO,CAIxE;AAED;;;;;;;GAOG;AACH,eAAO,MAAM,MAAM,aAA0C,CAAC;AAE9D;;GAEG;AACH,eAAO,MAAM,YAAY,MAAM,CAAC"}

package/dist/decode/edge-context.js

@@ -1,3 +1,10 @@
+/**
+ * What a rule can ask about the words around an edge.
+ *
+ * The lookups every edge rule shares: which word ends here, which starts here,
+ * and what part of speech the dictionary gives it.
+ */
+import { toCharacters } from "../script/characters.js";
 /**
  * The word ending at a position, longest first, or undefined.
  *
@@ -82,6 +89,24 @@ export function tagOf(context, word) {
         ? ""
         : (context.dictionary.lookup(word)?.partOfSpeech ?? "");
 }
+/**
+ * Whether a word of more than one character starts at a position.
+ *
+ * Every character is a word of its own, so a test that counted those would
+ * answer yes everywhere. What a rule asking this wants to know is whether some
+ * longer word has a claim on the character.
+ */
+export function startsLongerWord(context, at) {
+    return wordsStartingAt(context, at).some((word) => toCharacters(word).length > 1);
+}
+/**
+ * Whether a word of more than one character ends at a position.
+ *
+ * The mirror of {@link startsLongerWord}, and asked for the same reason.
+ */
+export function endsLongerWord(context, at) {
+    return wordsEndingAt(context, at).some((word) => toCharacters(word).length > 1);
+}
 /**
  * What follows a 教 that is teaching where no object does: aspect and the
  * complement marker, which only a verb takes.

package/dist/decode/edge-context.js.map

@@ -1 +1 @@
-{"version":3,"file":"edge-context.js","sourceRoot":"","sources":["../../src/decode/edge-context.ts"],"names":[],"mappings":"AAQA;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAC1B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,EAAE,GAAG,MAAM,CAAC;QACzB,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACzD,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAClD,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,aAAa,CAC3B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,EAAE,GAAG,MAAM,CAAC;QACzB,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACzD,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAClD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAClB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,mEAAmE;QACnE,sBAAsB;QACtB,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,GAAG,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAChE,IAAI,IAAI,KAAK,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACjE,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,eAAe,CAC7B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,GAAG,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAChE,IAAI,IAAI,KAAK,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACjE,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAClB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,KAAK,CAAC,OAAoB,EAAE,IAAwB;IAClE,OAAO,IAAI,KAAK,SAAS;QACvB,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,YAAY,IAAI,EAAE,CAAC,CAAC;AAC5D,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,CAAC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9D;;GAEG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,GAAG,CAAC"}
+{"version":3,"file":"edge-context.js","sourceRoot":"","sources":["../../src/decode/edge-context.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AACH,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAGvD;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAC1B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,EAAE,GAAG,MAAM,CAAC;QACzB,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACzD,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAClD,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,aAAa,CAC3B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,EAAE,GAAG,MAAM,CAAC;QACzB,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACzD,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAClD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAClB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,mEAAmE;QACnE,sBAAsB;QACtB,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,GAAG,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAChE,IAAI,IAAI,KAAK,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACjE,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,eAAe,CAC7B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,GAAG,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAChE,IAAI,IAAI,KAAK,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACjE,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAClB,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,KAAK,CAAC,OAAoB,EAAE,IAAwB;IAClE,OAAO,IAAI,KAAK,SAAS;QACvB,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,YAAY,IAAI,EAAE,CAAC,CAAC;AAC5D,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,gBAAgB,CAAC,OAAoB,EAAE,EAAU;IAC/D,OAAO,eAAe,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,IAAI,CACtC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CACxC,CAAC;AACJ,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,cAAc,CAAC,OAAoB,EAAE,EAAU;IAC7D,OAAO,aAAa,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,IAAI,CACpC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CACxC,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,CAAC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9D;;GAEG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,GAAG,CAAC"}

package/dist/decode/reading-rules.d.ts

@@ -1,6 +1,8 @@
 export { MODAL_DE, PARTICLE_DE, POTENTIAL_DE } from "./de-rule.js";
+export { TAXI_DI } from "./taxi-rule.js";
 export { EXPERIENTIAL_GUO } from "./guo-rule.js";
 export { TEACHING_JIAO } from "./jiao-rule.js";
+export { SEPARATED_COMPOUND } from "./split-rule.js";
 import { type EdgeRule } from "./rules.js";
 /**
  * 儿 does not stand on its own where the dictionary attests the 儿化.

package/dist/decode/reading-rules.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"reading-rules.d.ts","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAKA,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AACjD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,OAAO,EAEL,KAAK,QAAQ,EAGd,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;;;;GAmBG;AACH,eAAO,MAAM,cAAc,EAAE,QAY5B,CAAC;AAsCF;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,eAAe,EAAE,QAuB7B,CAAC;AAIF,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE;;GAEG;AACH,eAAO,MAAM,aAAa,EAAE,SAAS,QAAQ,EAU5C,CAAC"}
+{"version":3,"file":"reading-rules.d.ts","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AACjD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,kBAAkB,EAAE,MAAM,iBAAiB,CAAC;AAErD,OAAO,EAEL,KAAK,QAAQ,EAGd,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;;;;GAmBG;AACH,eAAO,MAAM,cAAc,EAAE,QAY5B,CAAC;AAsCF;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,eAAO,MAAM,eAAe,EAAE,QAuB7B,CAAC;AAIF,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE;;GAEG;AACH,eAAO,MAAM,aAAa,EAAE,SAAS,QAAQ,EAY5C,CAAC"}

package/dist/decode/reading-rules.js

@@ -1,10 +1,14 @@
 import { QUANTITY_CHARACTERS } from "../numerals/characters.js";
 import { MODAL_DE, PARTICLE_DE, POTENTIAL_DE } from "./de-rule.js";
+import { TAXI_DI } from "./taxi-rule.js";
 import { EXPERIENTIAL_GUO } from "./guo-rule.js";
 import { TEACHING_JIAO } from "./jiao-rule.js";
+import { SEPARATED_COMPOUND } from "./split-rule.js";
 export { MODAL_DE, PARTICLE_DE, POTENTIAL_DE } from "./de-rule.js";
+export { TAXI_DI } from "./taxi-rule.js";
 export { EXPERIENTIAL_GUO } from "./guo-rule.js";
 export { TEACHING_JIAO } from "./jiao-rule.js";
+export { SEPARATED_COMPOUND } from "./split-rule.js";
 import { toCharacters } from "../script/characters.js";
 import { tagOf, wordEndingAt, } from "./rules.js";
 /**
@@ -132,11 +136,13 @@ export const READING_RULES = [
     MODAL_DE,
     PARTICLE_DE,
     POTENTIAL_DE,
+    TAXI_DI,
     TEACHING_JIAO,
     ATTESTED_ERHUA,
     COUNTED_MEASURE,
     ADJECTIVAL_CHANG,
     PLAYING_TAN,
     EXPERIENTIAL_GUO,
+    SEPARATED_COMPOUND,
 ];
 //# sourceMappingURL=reading-rules.js.map

package/dist/decode/reading-rules.js.map

@@ -1 +1 @@
-{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AACjD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AACjD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AACvD,OAAO,EAGL,KAAK,EACL,YAAY,GACb,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,CAAC,MAAM,cAAc,GAAa;IACtC,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACzC,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,wEAAwE;QACxE,MAAM,QAAQ,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC;QACvD,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;QAC7D,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;GAKG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;;;;GAQG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAC/B,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,OAAO,IAAI,GAAG,CAAC,IAAI,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QACvE,IAAI,EAAE,CAAC;IACT,CAAC;IACD,IAAI,IAAI,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,KAAK,OAAO,EAAE,CAAC;QACpD,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IAC1E,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACxC,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAa;IACvC,IAAI,EAAE,iBAAiB;IACvB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,yEAAyE;QACzE,yEAAyE;QACzE,yEAAyE;QACzE,WAAW;QACX,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAC3C,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,OAAO,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1B,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,OAAO;YACtD,IAAI,CAAC,YAAY,KAAK,OAAO;YAC7B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB;IAChD,QAAQ;IACR,WAAW;IACX,YAAY;IACZ,aAAa;IACb,cAAc;IACd,eAAe;IACf,gBAAgB;IAChB,WAAW;IACX,gBAAgB;CACjB,CAAC"}
+{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AACjD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,kBAAkB,EAAE,MAAM,iBAAiB,CAAC;AAErD,OAAO,EAAE,QAAQ,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACnE,OAAO,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AACjD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,kBAAkB,EAAE,MAAM,iBAAiB,CAAC;AACrD,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AACvD,OAAO,EAGL,KAAK,EACL,YAAY,GACb,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,CAAC,MAAM,cAAc,GAAa;IACtC,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACzC,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,wEAAwE;QACxE,MAAM,QAAQ,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC;QACvD,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;QAC7D,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;GAKG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;;;;GAQG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAC/B,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,OAAO,IAAI,GAAG,CAAC,IAAI,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QACvE,IAAI,EAAE,CAAC;IACT,CAAC;IACD,IAAI,IAAI,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,KAAK,OAAO,EAAE,CAAC;QACpD,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IAC1E,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACxC,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAa;IACvC,IAAI,EAAE,iBAAiB;IACvB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,yEAAyE;QACzE,yEAAyE;QACzE,yEAAyE;QACzE,WAAW;QACX,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAC3C,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,OAAO,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1B,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,OAAO;YACtD,IAAI,CAAC,YAAY,KAAK,OAAO;YAC7B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB;IAChD,QAAQ;IACR,WAAW;IACX,YAAY;IACZ,OAAO;IACP,aAAa;IACb,cAAc;IACd,eAAe;IACf,gBAAgB;IAChB,WAAW;IACX,gBAAgB;IAChB,kBAAkB;CACnB,CAAC"}

package/dist/decode/rules.d.ts

@@ -1,6 +1,6 @@
 import type { Dictionary } from "../dictionary/dictionary.js";
 import type { Lattice, LatticeEdge } from "./lattice.js";
-export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordsEndingAt, wordStartingAt, wordsStartingAt, } from "./edge-context.js";
+export { ASPECT, endsLongerWord, PARTICLE_TAG, startsLongerWord, tagOf, wordEndingAt, wordsEndingAt, wordStartingAt, wordsStartingAt, } from "./edge-context.js";
 /**
  * What a rule can say about one edge.
  *

package/dist/decode/rules.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"rules.d.ts","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAEzD,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,aAAa,EACb,cAAc,EACd,eAAe,GAChB,MAAM,mBAAmB,CAAC;AAE3B;;;;;;;GAOG;AACH,MAAM,MAAM,WAAW,GAAG,MAAM,GAAG,QAAQ,GAAG,OAAO,CAAC;AAEtD;;;;;;GAMG;AACH,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,UAAU,EAAE,UAAU,CAAC;IAChC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,IAAI,EAAE,WAAW,CAAC;CAC5B;AAED;;;;;;;GAOG;AACH,MAAM,WAAW,QAAQ;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,UAAU,EAAE,CAAC,OAAO,EAAE,WAAW,KAAK,WAAW,CAAC;CAC5D;AA+DD;;;;;;;;;;;GAWG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,OAAO,EAChB,UAAU,EAAE,UAAU,EACtB,KAAK,EAAE,SAAS,QAAQ,EAAE,GACzB,OAAO,CAsCT"}
+{"version":3,"file":"rules.d.ts","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAEzD,OAAO,EACL,MAAM,EACN,cAAc,EACd,YAAY,EACZ,gBAAgB,EAChB,KAAK,EACL,YAAY,EACZ,aAAa,EACb,cAAc,EACd,eAAe,GAChB,MAAM,mBAAmB,CAAC;AAE3B;;;;;;;GAOG;AACH,MAAM,MAAM,WAAW,GAAG,MAAM,GAAG,QAAQ,GAAG,OAAO,CAAC;AAEtD;;;;;;GAMG;AACH,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,UAAU,EAAE,UAAU,CAAC;IAChC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,IAAI,EAAE,WAAW,CAAC;CAC5B;AAED;;;;;;;GAOG;AACH,MAAM,WAAW,QAAQ;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,UAAU,EAAE,CAAC,OAAO,EAAE,WAAW,KAAK,WAAW,CAAC;CAC5D;AA+DD;;;;;;;;;;;GAWG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,OAAO,EAChB,UAAU,EAAE,UAAU,EACtB,KAAK,EAAE,SAAS,QAAQ,EAAE,GACzB,OAAO,CAsCT"}

package/dist/decode/rules.js

@@ -1,4 +1,4 @@
-export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordsEndingAt, wordStartingAt, wordsStartingAt, } from "./edge-context.js";
+export { ASPECT, endsLongerWord, PARTICLE_TAG, startsLongerWord, tagOf, wordEndingAt, wordsEndingAt, wordStartingAt, wordsStartingAt, } from "./edge-context.js";
 /**
  * Whether every position can still be left, and the end still reached.
  *

package/dist/decode/rules.js.map

@@ -1 +1 @@
-{"version":3,"file":"rules.js","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAGA,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,aAAa,EACb,cAAc,EACd,eAAe,GAChB,MAAM,mBAAmB,CAAC;AAsC3B;;;;;;;GAOG;AACH,SAAS,aAAa,CAAC,OAAgB;IACrC,MAAM,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC;IACzC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,GAAG,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IAChE,OAAO,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;IACvB,KAAK,IAAI,EAAE,GAAG,MAAM,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC;QACxC,0EAA0E;QAC1E,OAAO,CAAC,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAC1C,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,KAAK,CACpC,CAAC;IACJ,CAAC;IACD,qDAAqD;IACrD,OAAO,OAAO,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC;AAC7B,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CACjB,KAA0B,EAC1B,OAAoB;IAEpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;YACvB,OAAO,OAAO,CAAC;QACjB,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,YAAY,CACnB,KAA6B,EAC7B,MAAgC,EAChC,SAA2B;IAE3B,MAAM,OAAO,GAAG,IAAI,GAAG,CACrB,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAChE,CAAC;IACF,KAAK,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAChD,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAgB,EAChB,UAAsB,EACtB,KAA0B;IAE1B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,GAAG,EAAe,CAAC;IACzC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAe,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,OAAO,GAAG,UAAU,CAAC,KAAK,EAAE;gBAChC,UAAU;gBACV,UAAU,EAAE,OAAO,CAAC,UAAU;gBAC9B,IAAI;aACL,CAAC,CAAC;YACH,IAAI,OAAO,KAAK,QAAQ,EAAE,CAAC;gBACzB,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACtB,CAAC;iBAAM,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;gBAC/B,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,4EAA4E;IAC5E,kEAAkE;IAClE,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,CAAC,CAAC;IACzC,CAAC;IAED,IAAI,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,KAAK,GAAY;QACrB,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CACjC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAC7C;KACF,CAAC;IACF,OAAO,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC;AAChD,CAAC"}
+{"version":3,"file":"rules.js","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAGA,OAAO,EACL,MAAM,EACN,cAAc,EACd,YAAY,EACZ,gBAAgB,EAChB,KAAK,EACL,YAAY,EACZ,aAAa,EACb,cAAc,EACd,eAAe,GAChB,MAAM,mBAAmB,CAAC;AAsC3B;;;;;;;GAOG;AACH,SAAS,aAAa,CAAC,OAAgB;IACrC,MAAM,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC;IACzC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,GAAG,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IAChE,OAAO,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;IACvB,KAAK,IAAI,EAAE,GAAG,MAAM,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC;QACxC,0EAA0E;QAC1E,OAAO,CAAC,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAC1C,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,KAAK,CACpC,CAAC;IACJ,CAAC;IACD,qDAAqD;IACrD,OAAO,OAAO,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC;AAC7B,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CACjB,KAA0B,EAC1B,OAAoB;IAEpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;YACvB,OAAO,OAAO,CAAC;QACjB,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,YAAY,CACnB,KAA6B,EAC7B,MAAgC,EAChC,SAA2B;IAE3B,MAAM,OAAO,GAAG,IAAI,GAAG,CACrB,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAChE,CAAC;IACF,KAAK,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAChD,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAgB,EAChB,UAAsB,EACtB,KAA0B;IAE1B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,GAAG,EAAe,CAAC;IACzC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAe,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,OAAO,GAAG,UAAU,CAAC,KAAK,EAAE;gBAChC,UAAU;gBACV,UAAU,EAAE,OAAO,CAAC,UAAU;gBAC9B,IAAI;aACL,CAAC,CAAC;YACH,IAAI,OAAO,KAAK,QAAQ,EAAE,CAAC;gBACzB,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACtB,CAAC;iBAAM,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;gBAC/B,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,4EAA4E;IAC5E,kEAAkE;IAClE,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,CAAC,CAAC;IACzC,CAAC;IAED,IAAI,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,KAAK,GAAY;QACrB,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CACjC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAC7C;KACF,CAAC;IACF,OAAO,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC;AAChD,CAAC"}

package/dist/decode/split-rule.d.ts

@@ -0,0 +1,43 @@
+import { type EdgeRule } from "./rules.js";
+/**
+ * A 离合词 keeps its reading when a 量词 is pushed into the middle of it.
+ *
+ * 请假 is `qǐngjià` and 请个假 came out `qǐng gè jiǎ`, because the word the
+ * reading hangs on is no longer there to be matched and each half falls back on
+ * its character's default. 睡个觉 went to `jué`, 弹个琴 to `dàn`, 教个书 to
+ * `jiào` and 有个空 to `kōng`, all of them the reading the compound exists to
+ * rule out. Nothing here is new about the characters. The dictionary already
+ * holds `jià` as an alternate of 假 and already states that 请假 is `qǐng jià`.
+ * What was missing is that the second fact reaches the first across a 个.
+ *
+ * A 离合词 is a verb and its object written as one word, so the halves come
+ * apart and take an infix between them the way no other compound does. The
+ * classifier is the commonest infix, and that is what makes the shape
+ * decidable. See {@link separatedSyllableAt} for the three conditions and
+ * {@link CLASSIFIER} for why no other 量词 is in.
+ *
+ * Both halves are settled, because either can be the polyphone. 请个假 and
+ * 睡个觉 misread the tail, 弹个琴 and 教个书 the head.
+ *
+ * Measured over the same 88,866 lines of Tatoeba and zh.wikipedia the other
+ * rules were sized against, 2,018 head-个-tail shapes have a dictionary pair
+ * behind them. The conditions admit 3 and all 3 are corrections (请个假 to
+ * `jià`, 打个折 to `zhé` and 睡個覺 to `jiào`). The 66 they decline are pairs the
+ * index happens to hold (两个都 off 两都, 一个和 off 一和, 那个是 off 那是,
+ * 一个只 off 一只), where no word was ever split in two, and the verb-head
+ * condition alone accounts for 44 of them.
+ *
+ * The one real firing given up is 瞅个空 for `kòng`, which 瞅空 would carry if
+ * an untagged pair were allowed to. It goes with 是个甚, and that trade is the
+ * right way round, since the corpus has one of each and admitting untagged
+ * pairs opens the rule to 720,000 keys nobody counted.
+ *
+ * The measurement is thin because the shape is rare in written text, so the
+ * conditions were sized against the dictionary as well. 2,994 tagged pairs with
+ * a verb in front hold a half whose reading differs from its character's
+ * default, and they are led by 认为, 获得, 关系 and 告诉, none of which any 个
+ * can be pushed into. The rule never reaches them, since it fires on a 个 that
+ * is written rather than on a pair that exists.
+ */
+export declare const SEPARATED_COMPOUND: EdgeRule;
+//# sourceMappingURL=split-rule.d.ts.map

package/dist/decode/split-rule.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"split-rule.d.ts","sourceRoot":"","sources":["../../src/decode/split-rule.ts"],"names":[],"mappings":"AASA,OAAO,EAEL,KAAK,QAAQ,EAGd,MAAM,YAAY,CAAC;AA+HpB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AACH,eAAO,MAAM,kBAAkB,EAAE,QAehC,CAAC"}

package/dist/decode/split-rule.js

@@ -0,0 +1,174 @@
+/**
+ * How a 离合词 is read once a 量词 has been pushed into the middle of it.
+ *
+ * 请假 is one word and carries one reading. 请个假 is the same word with a
+ * classifier inside it, and the halves are left standing on their own.
+ */
+import { isSameSyllable } from "../dictionary/entry.js";
+import { toCharacters } from "../script/characters.js";
+import { tagOf, wordsStartingAt, } from "./rules.js";
+/**
+ * The classifier that separates the halves, in both scripts.
+ *
+ * 个 alone, and the measurement below is the reason. Every other 量词 was tried
+ * and the noise buried the signal. The dictionary tags 道, 名, 家, 子, 口 and 集
+ * `q` too, and those spend almost all of their time as ordinary morphemes, so
+ * 知道了 read its 了 as `liǎo` off 知了 and 有道理 its 理 as `lǐ` off 有理.
+ * Widening to the whole `q` tag carries the rule from 3 firings to 109, of
+ * which 46 are that one 知道了.
+ *
+ * 下 and 回 are outside it for a different reason. Both separate a compound
+ * readily enough, but both also head a directional complement, and 坐下来,
+ * 拿回来 and 生下来 all end in a 来 that is neutral where the compound behind
+ * them says `lái`. Adding the pair costs 34 wrong firings to gain none.
+ */
+const CLASSIFIER = new Set(["个", "個"]);
+/**
+ * The tag prefix jieba gives a verb.
+ *
+ * A prefix, so that `vn` and `vd` are in. 有空 is tagged `vn`.
+ */
+const VERB_TAG = "v";
+/**
+ * The two halves a classifier separates at a position, or undefined.
+ *
+ * The position is one half of the pair and the classifier says which. A 个 in
+ * front of it makes it the tail, and a 个 behind it makes it the head. Both are
+ * wanted, because either half can be the polyphone. 请个假 misreads the 假 and
+ * 弹个琴 the 弹.
+ */
+function separatedAt(context, at) {
+    const { characters } = context;
+    if (CLASSIFIER.has(characters[at + 1] ?? "") && at + 2 < characters.length) {
+        return {
+            /* c8 ignore next -- both positions are inside the run just measured */
+            word: `${characters[at] ?? ""}${characters[at + 2] ?? ""}`,
+            offset: 0,
+        };
+    }
+    if (CLASSIFIER.has(characters[at - 1] ?? "") && at >= 2) {
+        return {
+            /* c8 ignore next -- both positions are inside the run just measured */
+            word: `${characters[at - 2] ?? ""}${characters[at] ?? ""}`,
+            offset: 1,
+        };
+    }
+    return undefined;
+}
+/**
+ * Whether a tagged word of more than one character starts at a position.
+ *
+ * The guard that keeps the rule off a 个 that is counting the word after it.
+ * 我看见那儿有个奇怪的女人 has 有奇 behind it, a key meaning *and a fraction
+ * more*, and forcing its `jī` onto a 奇 that opens 奇怪 is the one wrong firing
+ * the corpus offers. A tag is asked for and not merely a key, because the full
+ * tier holds 720,000 of them and 打个折么 finds 折么 that way. The tag is the
+ * record that somebody counted the word.
+ */
+function opensTaggedWord(context, at) {
+    return wordsStartingAt(context, at).some((word) => toCharacters(word).length > 1 && tagOf(context, word) !== "");
+}
+/**
+ * The syllable a separated compound gives one of its halves, or undefined.
+ *
+ * Three things have to hold, and each one is carrying its weight over the
+ * corpus. The pair has to be a word the dictionary tags, since an untagged key
+ * is a reading somebody recorded rather than a word anybody counted, and
+ * 是个甚么 finds 是甚 that way and would read its 甚 as `shèn`. The character in
+ * front has to be a verb. That is what separates 请个假 from 一个只, 两个都 and
+ * 八个行, where the 个 counts the numeral in front of it and the pair behind is
+ * an accident. And the far half has to stand clear of a tagged word of its own,
+ * for the reason {@link opensTaggedWord} gives.
+ */
+function separatedSyllableAt(context, at) {
+    const separated = separatedAt(context, at);
+    if (separated === undefined) {
+        return undefined;
+    }
+    const { word, offset } = separated;
+    const entry = context.dictionary.lookup(word);
+    if (entry?.reading.length !== 2 || entry.partOfSpeech === "") {
+        return undefined;
+    }
+    const [head = ""] = toCharacters(word);
+    if (!tagOf(context, head).startsWith(VERB_TAG)) {
+        return undefined;
+    }
+    const tail = offset === 1 ? at : at + 2;
+    return opensTaggedWord(context, tail) ? undefined : entry.reading[offset];
+}
+/**
+ * Whether a pair the dictionary holds carries the unseparated reading across.
+ *
+ * Forcing the single-character edge settles nothing on its own, for the reason
+ * {@link TEACHING_JIAO} gives. A reading spanning two characters brings its own
+ * half with it. 能给我打个折么 read `shémǒ` off 折么, a key the dictionary holds
+ * with no part of speech and at the cost of a word nothing has ever counted.
+ * A *tagged* word is left alone, which is every one that matters.
+ */
+function carriesReading(context, edge) {
+    if (edge.partOfSpeech !== "" || edge.reading.length !== edge.to - edge.from) {
+        return false;
+    }
+    return edge.reading.some((syllable, offset) => {
+        const said = separatedSyllableAt(context, edge.from + offset);
+        return said !== undefined && !isSameSyllable(syllable, said);
+    });
+}
+/**
+ * A 离合词 keeps its reading when a 量词 is pushed into the middle of it.
+ *
+ * 请假 is `qǐngjià` and 请个假 came out `qǐng gè jiǎ`, because the word the
+ * reading hangs on is no longer there to be matched and each half falls back on
+ * its character's default. 睡个觉 went to `jué`, 弹个琴 to `dàn`, 教个书 to
+ * `jiào` and 有个空 to `kōng`, all of them the reading the compound exists to
+ * rule out. Nothing here is new about the characters. The dictionary already
+ * holds `jià` as an alternate of 假 and already states that 请假 is `qǐng jià`.
+ * What was missing is that the second fact reaches the first across a 个.
+ *
+ * A 离合词 is a verb and its object written as one word, so the halves come
+ * apart and take an infix between them the way no other compound does. The
+ * classifier is the commonest infix, and that is what makes the shape
+ * decidable. See {@link separatedSyllableAt} for the three conditions and
+ * {@link CLASSIFIER} for why no other 量词 is in.
+ *
+ * Both halves are settled, because either can be the polyphone. 请个假 and
+ * 睡个觉 misread the tail, 弹个琴 and 教个书 the head.
+ *
+ * Measured over the same 88,866 lines of Tatoeba and zh.wikipedia the other
+ * rules were sized against, 2,018 head-个-tail shapes have a dictionary pair
+ * behind them. The conditions admit 3 and all 3 are corrections (请个假 to
+ * `jià`, 打个折 to `zhé` and 睡個覺 to `jiào`). The 66 they decline are pairs the
+ * index happens to hold (两个都 off 两都, 一个和 off 一和, 那个是 off 那是,
+ * 一个只 off 一只), where no word was ever split in two, and the verb-head
+ * condition alone accounts for 44 of them.
+ *
+ * The one real firing given up is 瞅个空 for `kòng`, which 瞅空 would carry if
+ * an untagged pair were allowed to. It goes with 是个甚, and that trade is the
+ * right way round, since the corpus has one of each and admitting untagged
+ * pairs opens the rule to 720,000 keys nobody counted.
+ *
+ * The measurement is thin because the shape is rare in written text, so the
+ * conditions were sized against the dictionary as well. 2,994 tagged pairs with
+ * a verb in front hold a half whose reading differs from its character's
+ * default, and they are led by 认为, 获得, 关系 and 告诉, none of which any 个
+ * can be pushed into. The rule never reaches them, since it fires on a 个 that
+ * is written rather than on a pair that exists.
+ */
+export const SEPARATED_COMPOUND = {
+    name: "separated-compound",
+    verdictFor: (context) => {
+        const { edge } = context;
+        if (edge.to - edge.from !== 1 || edge.reading.length !== 1) {
+            return carriesReading(context, edge) ? "forbid" : "keep";
+        }
+        const said = separatedSyllableAt(context, edge.from);
+        const here = edge.reading[0];
+        /* c8 ignore next -- a one-syllable reading has a syllable at zero */
+        if (said === undefined || here === undefined) {
+            return "keep";
+        }
+        return isSameSyllable(here, said) ? "force" : "keep";
+    },
+};
+//# sourceMappingURL=split-rule.js.map

package/dist/decode/split-rule.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"split-rule.js","sourceRoot":"","sources":["../../src/decode/split-rule.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AACH,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAEvD,OAAO,EAGL,KAAK,EACL,eAAe,GAChB,MAAM,YAAY,CAAC;AAEpB;;;;;;;;;;;;;;GAcG;AACH,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEvC;;;;GAIG;AACH,MAAM,QAAQ,GAAG,GAAG,CAAC;AAErB;;;;;;;GAOG;AACH,SAAS,WAAW,CAClB,OAAoB,EACpB,EAAU;IAEV,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAC/B,IAAI,UAAU,CAAC,GAAG,CAAC,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,GAAG,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC;QAC3E,OAAO;YACL,uEAAuE;YACvE,IAAI,EAAE,GAAG,UAAU,CAAC,EAAE,CAAC,IAAI,EAAE,GAAG,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;YAC1D,MAAM,EAAE,CAAC;SACV,CAAC;IACJ,CAAC;IACD,IAAI,UAAU,CAAC,GAAG,CAAC,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC;QACxD,OAAO;YACL,uEAAuE;YACvE,IAAI,EAAE,GAAG,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,UAAU,CAAC,EAAE,CAAC,IAAI,EAAE,EAAE;YAC1D,MAAM,EAAE,CAAC;SACV,CAAC;IACJ,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;;;;;GASG;AACH,SAAS,eAAe,CAAC,OAAoB,EAAE,EAAU;IACvD,OAAO,eAAe,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,IAAI,CACtC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,EAAE,CACvE,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;GAWG;AACH,SAAS,mBAAmB,CAC1B,OAAoB,EACpB,EAAU;IAEV,MAAM,SAAS,GAAG,WAAW,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IAC3C,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC5B,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,SAAS,CAAC;IACnC,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAC9C,IAAI,KAAK,EAAE,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,YAAY,KAAK,EAAE,EAAE,CAAC;QAC7D,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,CAAC,IAAI,GAAG,EAAE,CAAC,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IACvC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC/C,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IACxC,OAAO,eAAe,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;AAC5E,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,cAAc,CACrB,OAAoB,EACpB,IAAyB;IAEzB,IAAI,IAAI,CAAC,YAAY,KAAK,EAAE,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;QAC5E,OAAO,KAAK,CAAC;IACf,CAAC;IACD,OAAO,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,MAAM,EAAE,EAAE;QAC5C,MAAM,IAAI,GAAG,mBAAmB,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,GAAG,MAAM,CAAC,CAAC;QAC9D,OAAO,IAAI,KAAK,SAAS,IAAI,CAAC,cAAc,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;IAC/D,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAa;IAC1C,IAAI,EAAE,oBAAoB;IAC1B,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,KAAK,CAAC,IAAI,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC3D,OAAO,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;QAC3D,CAAC;QACD,MAAM,IAAI,GAAG,mBAAmB,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QACrD,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QAC7B,qEAAqE;QACrE,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7C,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,cAAc,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC;IACvD,CAAC;CACF,CAAC"}

package/dist/decode/taxi-rule.d.ts

@@ -0,0 +1,45 @@
+import { type EdgeRule } from "./rules.js";
+/**
+ * 的 read `dī` where the taxi word holding it stands, and written whole there.
+ *
+ * The taxi vocabulary is six keys (的士, 的哥, 的姐, 打的, 面的 and 摩的). The
+ * `dī` in them is the only `dī` the dictionary holds, and that reading names
+ * the set. All six also spell a modifier, a particle 的 and a head, much the
+ * commoner reading of the same characters, and the decode had it both ways at
+ * once. 我的哥哥们 came out `wǒ dī gēgē men` on the strength of 的哥, while
+ * 的士 came out `dī shì` on the strength of nothing at all.
+ *
+ * What decides it is what follows the 的, and where the key sits around the 的
+ * decides which question that is.
+ *
+ * A key **beginning** with the 的 offers its own second character as the head.
+ * A longer word starting there is the better claim on it. 哥哥 takes the 哥 of
+ * 我的哥哥们, 姐姐 the 姐 of 我的姐姐 and 士兵 the 士 of 我的士兵. Nothing
+ * starts at the 哥 of 那个的哥 or the 士 of 我要一辆的士, and both keys stand.
+ *
+ * A key **ending** with the 的 puts the head outside itself, and
+ * {@link AFTER_PARTICLE} is what one looks like. The other side is asked as
+ * well. A longer word ending at the 的 owns the character in front of it, and
+ * that is what 上面的, 方面的 and 表面的 are. Without that test 26 more runs of
+ * the corpus change and every one of them is wrong.
+ *
+ * Where the word does stand the bare 的 is taken out from under it. That is the
+ * spacing half of the same claim. 的 sits in the cheapest band the dictionary
+ * holds, at 5.62 against a 4.62 floor. A two-character key spanning one is
+ * therefore weighed against a split starting several buckets ahead of an
+ * ordinary word's, and 的士 came apart at 16.62 against 16.24 for the split
+ * where 的哥 at 14.62 held. Forbidding the split is what reaches spacing at
+ * all, since forcing an edge only settles the rivals covering the same
+ * characters.
+ *
+ * Measured over the 88,866 lines of Tatoeba and zh.wikipedia, 21 的 read `dī`
+ * and 9 of them were the particle. This changes 17 runs. Six are readings it
+ * corrects (所打的仗, 谁打的电话, 打的不是很好, 我的哥哥们, 我打的这辆出租车,
+ * 见过面的那个人), ten are taxi words that were split and now hold together,
+ * and the last is 用一元硬币来打的, which joins up while staying wrong. That
+ * one closes a 是⋯的 whose 是 sits in the clause before it, and no local
+ * evidence says so. The two others left are the 的哥里 of a transliterated
+ * Gori. The gold corpus holds at 3 misses and CPP at 91.49%.
+ */
+export declare const TAXI_DI: EdgeRule;
+//# sourceMappingURL=taxi-rule.d.ts.map

package/dist/decode/taxi-rule.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"taxi-rule.d.ts","sourceRoot":"","sources":["../../src/decode/taxi-rule.ts"],"names":[],"mappings":"AAQA,OAAO,EAEL,KAAK,QAAQ,EAKd,MAAM,YAAY,CAAC;AA2EpB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAyCG;AACH,eAAO,MAAM,OAAO,EAAE,QAYrB,CAAC"}